NATs at a glance Free

N-terminal acetyltransferases (NATs) catalyse the transfer of an acetyl group (-COCH₃) to the very N-terminal amino group (-NH₃⁺) of proteins. This causes an increase in the hydrophobicity of the protein terminus and can affect several aspects of protein functionality. Consequently, impairment of NAT activity triggers cellular and physiological effects that reflect malfunction of NAT substrates with an unmodified N terminus. Each human NAT catalyses N-terminal acetylation (Nt-acetylation) of distinct substrates, and together the NATs ensure that 80% of the proteome receives this modification (Arnesen et al., 2009b; Aksnes et al., 2015). The main features of the NAT machinery and the N-terminal acetylome are conserved among eukaryotes (Giglione and Meinnel, 2021; Goetze et al., 2009; Arnesen et al., 2009b). Seven human NATs are known (see poster). The bulk of cellular Nt-acetylation events in humans are driven by the workhorses NatA, NatB and NatC, and to some extent NatE, whereas NatD and NatH have highly selective substrate specificity, and NatF targets a subgroup of membrane proteins (Aksnes et al., 2019). Which NAT acts on a given protein is largely determined by the N-terminal sequence of the substrate. Besides covering separate substrate pools, the NATs are also distinguished by several features as highlighted below. This Cell Science at a Glance article and the accompanying poster provide an introduction to Nt-acetylation and the NAT enzymes catalysing this modification, including the cellular and physiological consequences of NAT deficiency. Here, we mainly focus on the human NATs, but we also include insights based on findings from a variety of model organisms.

The first identified eukaryotic NAT, NatA, is also the most important NAT in terms of the number of substrates. NatA provides co-translational Nt-acetylation of proteins with N termini that have been processed by methionine aminopeptidases (MetAPs). This set of proteins comprises those with sequences starting (M)S, (M)A, (M)T, (M)G and (M)V, and probably also includes those starting (M)C (parentheses indicate the cleaved N-terminal methionine). NatA is the only NAT broadly targeting such a large group of protein N termini, which constitutes ∼40% of the human proteome (Arnesen et al., 2009b) (see poster).

NatA contains two main subunits, the catalytic NAA10 and the ribosome-anchoring NAA15 (Arnesen et al., 2005; Mullen et al., 1989; Gautschi et al., 2003). NAA15 wraps around NAA10 and modulates its catalytic cleft to accommodate the co-translational substrates of NatA (Liszczak et al., 2013). In addition, the chaperone-like huntingtin-interacting protein K (HYPK) stably associates with the NatA complex, and although it displays inhibitory activity in vitro (Gottlieb and Marmorstein, 2018; Weyer et al., 2017), HYPK is believed to promote NatA-mediated Nt-acetylation in vivo (Arnesen et al., 2010; Miklankova et al., 2022). A cryo-electron microscopy (cryo-EM) structure of the ribosome–NatA complex has shown that NatA is dynamically positioned directly underneath the ribosomal exit tunnel via a unique type of ribosome interaction that involves ribosomal RNA expansion segments (Knorr et al., 2019). NAA15 resides in the cytosol, but NAA10 localizes to both the cytosol and nucleus (Arnesen et al., 2005), thus suggesting that NAA10 exerts NatA-independent functions; both non-catalytic regulation of proteins and lysine acetylation have been proposed (Aksnes et al., 2019). In human cells, there are paralogues of NAA15 and NAA10: NAA16 (hNAT2) and NAA11 (ARD2), respectively (Arnesen et al., 2009a, 2006b). Although these proteins appear to be less abundant than NAA15 and NAA10 in most cells, they are likely to contribute to NatA-mediated Nt-acetylation.

NatA is essential for life in multicellular eukaryotes (Van Damme et al., 2011a; Wang et al., 2010; Ree et al., 2015; Ingram et al., 2000; Linster et al., 2015; Sonnichsen et al., 2005). As a reflection of the large pool of NatA substrates, the cellular roles of NatA are manifold. In human cells, loss of NatA activity can induce growth arrest and p53 (TP53)-dependent apoptosis, indicating a role in cell proliferation and survival (Arnesen et al., 2006c; Gromyko et al., 2010; Myklebust et al., 2015). One important function of NatA is to protect proteins from degradation, since unacetylated NatA substrates are recognized by the inhibitor of apoptosis protein (IAP) family of E3 ubiquitin ligases and targeted for degradation (Mueller et al., 2021). These IAPs are required for the apoptosis resulting from NatA knockdown, suggesting the involvement of protein degradation of unacetylated NatA substrates in this process (see poster) (Mueller et al., 2021). NatA acting as a protein stabilizer might be a conserved effect among the yeast, plant and human proteomes (see poster) (Mueller et al., 2021; Linster et al., 2022; Guzman et al., 2022 preprint). In contrast, Nt-acetylation of certain yeast NatA substrates can act as an N-terminal protein degradation signal (N-degron; Hwang et al., 2010).

In 2011, the first pathogenic NAT variant causing human disease was identified in NAA10. This NAA10 missense variant was present in eight boys with global developmental delay, hypotonia, aged appearance, craniofacial anomalies and cardiac arrhythmias, who died within two years of life (Rope et al., 2011). In the past decade, a number of males and females with additional pathogenic variants in NAA10, as well as in NAA15, have been identified (McTiernan et al., 2022; Cheng et al., 2020; Ward et al., 2021; Cheng et al., 2018; Lyon et al., 2023). Within this group, the shared phenotypes are developmental delay, intellectual disability, congenital heart disease and autism spectrum disorder. These clinical features might at least in part be caused by impaired NatA activity, which is likely to have pleiotropic effects due to the numerous substrates of NatA (McTiernan et al., 2022; Cheng et al., 2020, 2018; Ward et al., 2021; Myklebust et al., 2015). NatA also plays a role in cancer, where NAA10 appears to have both pro- and anti-tumorigenic features depending on the condition (see Box 1).

With a proteome coverage of ∼20%, NatB is the second major contributor to the N-terminal acetylome (Nt-acetylome). In contrast to NatA, NatB only targets substrates starting with a methionine. As a further determinant for NatB processing, the methionine needs to be followed by an acidic or amidic residue. Thus, proteins with N termini modified by NatB start with MD, ME, MN and MQ (Van Damme et al., 2012). Proteins with this type of N-terminal sequence exist in cells almost exclusively in the N-terminally acetylated (Nt-acetylated) state. This is a unique feature of NatB activity; for other multi-target NATs, it is quite common that a specific substrate exists as both Nt-acetylated and non-Nt-acetylated versions within its protein pool. NatB substrates are, however, very highly Nt-acetylated (typically 90–100%), indicating thorough coverage by NatB (Aksnes et al., 2016).

NatB has two subunits, the catalytic NAA20 and the ribosome-anchoring NAA25 (Polevoda et al., 2003; Starheim et al., 2008) (see poster), and is thus considered to be a co-translational NAT. Overall, the structure of the NatB complex (Deng et al., 2020b) is very similar to that of NatA (Liszczak et al., 2013), in that the large ribosome-binding subunit has a series of α-helices forming a long sheet from the N terminus to the C terminus, the core region of which wraps around the catalytic subunit (see poster). However, the orientation of the C terminus of the ribosome-binding subunit is differentially positioned in NatA and NatB (Deng et al., 2020b).

Cellular phenotypes resulting from NatB knockdown include altered cytoskeleton morphology, delocalization of myosin II, and a decrease in the number and size of focal adhesions, along with reduced cellular motility (Van Damme et al., 2012). These phenotypes can be reversed by overexpression of tropomyosin 1 (TPM1, which has an MD sequence at the N terminus); thus, similar to the situation in yeast (Polevoda et al., 2003; Singer and Shaw, 2003), a defective TPM1 in a non-Nt-acetylated state has been suggested to be the main determinant of these actomyosin phenotypes (see poster) (Van Damme et al., 2012). The results of knockdown experiments using cultured cancer cell lines also suggest that NatB is important for cell cycle progression (Starheim et al., 2008; Ametzazurra et al., 2008). More recently, CRISPR/Cas9-mediated knockout of NAA20 and NAA25 has revealed that NatB-mediated Nt-acetylation is required for the mechanism by which influenza A virus suppresses host gene expression, which is mediated by Nt-acetylation-dependent functionality of an endonuclease active site (Oishi et al., 2018).

Recently, the first pathogenic NAA20 variants have been identified. In agreement with the potentially large number of substrates affected by NatB impairment, individuals with these NAA20 variants present with pleiotropic phenotypes, including developmental delay, intellectual disability and microcephaly (Morrison et al., 2021; D'Onofrio et al., 2023). Among proteins matching the substrate specificity of NatB is the Parkinson's disease-related α-synuclein (αSyn), which has an MD sequence at the N terminus. Although direct evidence for NatB-mediated Nt-acetylation of αSyn is lacking, the N-terminal MD sequence is a strong indicator that αSyn is a NatB substrate. Furthermore, a cryo-EM structure of NatB bound to a coenzyme A (CoA)–αSyn conjugate has been reported (Deng et al., 2020b). In vitro experiments using non-Nt-acetylated αSyn have shown that this version has increased aggregation propensity (Kang et al., 2012; Bu et al., 2017; Bell et al., 2022). It is therefore possible that Nt-acetylation has a protective effect against αSyn oligomerization in vivo, although this has not been studied thus far. NatB has also been linked to cancer (Box 1).

Similarly to NatA and NatB, NatC consists of a catalytic subunit, NAA30, and a ribosome-anchoring subunit, NAA35, but it additionally contains NAA38, which is an obligatory subunit with a less well-described function (Polevoda and Sherman, 2001). The overall structure of NatC is different from that of NatA and NatB, with the ribosome-binding subunit only partially wrapping around the catalytic subunit and with a different inter-subunit orientation. Recent comparisons of the cryo-EM structures of human NatC complexes with and without NAA38 have revealed that NAA38 increases the thermostability and broadens the substrate-specificity profile of NatC by affecting structural features of both NAA30 and NAA35 (Deng et al., 2023, 2021a; Grunwald et al., 2020).

NatC N-terminally acetylates the subset of methionine-starting proteins not targeted by NatB. Thus, NatC modifies the N terminus of proteins starting with ML, MI, MF, MV, MY, MW, MH or MK sequences (Van Damme et al., 2016, 2023). It is noteworthy that proteins containing this type of N terminus are less frequently Nt-acetylated, and the target substrate proteins are typically Nt-acetylated in only a fraction of the overall molecular pool. It is therefore not straightforward to predict substrates of NatC-mediated Nt-acetylation. However, N-terminal proteomic analysis of NatC-lacking cells has identified 46 human (Van Damme et al., 2016) and 57 yeast (Van Damme et al., 2023) NatC substrates.

Cellular phenotypes resulting from NAA30 deficiency point to a role of NatC in mitochondrial function and cell survival. Knockdown of NAA30 leads to apoptosis, involving stabilization of p53 and induction of downstream proapoptotic genes (Starheim et al., 2009). Mitochondrial fragmentation and loss of mitochondrial membrane potential, with accompanying reduced expression levels of mitochondrial matrix proteins, are also observed upon NAA30 knockdown, occurring independently of p53 pathways (Van Damme et al., 2016). A role of NatC in mitochondrial function is also seen in Saccharomyces cerevisiae (Van Damme et al., 2023), and findings from Caenorhabditis elegans point to a regulatory role in response to nutrient availability or stressors, in which NatC up- and down-regulation potentially coordinates the balance between growth and development versus stress responses and energy-saving quiescence (Malinow et al., 2022; Warnhoff et al., 2014). A key role for NatC-mediated Nt-acetylation is to shield human proteins from degradation, as unacetylated NatC substrates are recognized by specific E3 ubiquitin ligases (UBR4–KCMF1, UBR1 and UBR2) and targeted for proteasomal degradation (see poster) (Varland et al., 2022 preprint). NatC-knockout phenotypes are reversed in the absence of these E3 ubiquitin ligases (Varland et al., 2022 preprint), demonstrating the central cellular interplay between these pathways: Nt-acetylation and stability versus degradation resulting from exposure of an N-degron. Particularly important targets of NatC-mediated Nt-acetylation appear to be the cullin E2 NEDD8-conjugating enzymes UBE2M and UBE2F; their Nt-acetylation is not only important for preventing their protein degradation, but also as an avidity enhancer, by allowing burial of their N termini into hydrophobic pockets of their cognate E3 ligases (Scott et al., 2011; Monda et al., 2013). The importance of Nt-acetylation of UBE2M and UBE2F is emphasized by the finding that the longevity and motility defects of Drosophila melanogaster Naa30A deletion mutants could partially be complemented by overexpression of UbcE2M, the only Drosophila UBE2M homologue (Varland et al., 2022 preprint). Consistent with a role for NatC in cell energy regulation and survival, NAA30 expression is frequently upregulated in cancer (Box 1).

NatD is a highly selective NAT, targeting only histones H2A and H4, as well as a few other substrates with similar N-terminal sequences (Song et al., 2003; Hole et al., 2011; Jonckheere and Van Damme, 2021). An additional particularity of NatD is that it is believed to be composed solely of the catalytic unit NAA40, since no stable binding partner has been found (Aksnes et al., 2019), and it has been proposed that NAA40 might itself bind to the ribosome via an extension not present in other NATs (Magin et al., 2015) (see poster).

Since NatD targets histones, its cellular function resembles that of histone acetyltransferases (HATs; see poster) (Demetriadou et al., 2020; Constantinou et al., 2023). HATs acetylate lysine sidechains on histones, with implications for gene regulation, and such histone modifiers often act in response to metabolic signals (Li et al., 2018). Interestingly, however, NatD is one of the first epigenetic modifiers found to act upstream of a metabolic pathway. Depletion of NAA40 in murine hepatocytes leads to an increase in intracellular acetyl CoA (AcCoA) levels, with further impacts on lipid synthesis and insulin signalling; and correspondingly, NAA40 expression is inversely correlated with insulin-resistant traits in individuals with obesity (Charidemou et al., 2022).

NatD has been reported to have an oncogenic role in various cancer types (Koufaris and Kirmizis, 2021; Demetriadou et al., 2019; Ju et al., 2017) (Box 1), and reported cellular phenotypes concern cancerous processes. Knockdown of NAA40 in colorectal cancer cells causes decreased survival as a consequence of increased p53-independent apoptosis, which occurs via the mitochondrial caspase-9-mediated apoptotic pathway (Pavlou and Kirmizis, 2016). NAA40 is also involved in chemoresistance through regulation of one-carbon metabolism – a second indication that NatD is a metabolic regulator (Demetriadou et al., 2022). Interestingly, since NatD has high selectivity towards the N-terminal SGRGK sequence, which is common to both H2A and H4, oncohistone mutations within this sequence suppress NatD-mediated Nt-acetylation; however, it is unclear whether this contributes to the oncogenic effect (Ho and Huang, 2022).

NatE activity results from the binding of catalytically active NAA50 to NatA subunits. NAA50 forms evolutionarily conserved contacts with both NAA10 and NAA15, which promotes NatE activity (Deng et al., 2019; Gautschi et al., 2003), whereas binding to HYPK has an inhibitory effect on NatE in vitro (Deng et al., 2020a). Although human NAA50 interacts with NAA10 and NAA15 (Arnesen et al., 2006a), a major fraction of NAA50 appears to act independently of the NatA complex (Hou et al., 2007). Despite having this physical connection with NatA, the substrate specificity of NatE overlaps with that of NatC (and NatF), as NatE targets methionine when it is followed by a hydrophobic or amphipathic residue (Van Damme et al., 2011a; Evjenth et al., 2009), as well as potentially competing with MetAP and NatA for some N termini (Van Damme et al., 2015). Thus, NatE has a mix of properties that overlap with different NATs (see poster).

The function of NAA50 is quite enigmatic. Although the catalytic function of the S. cerevisiae NAA50 homologue is questioned (Van Damme et al., 2015; Deng et al., 2019), NatE activity has been demonstrated in the plant Arabidopsis thaliana, where it is important for development and regulation of stress responses (Armbruster et al., 2020), as well as pollen development (Feng et al., 2022) (see poster for an overview of NAT subunits in model species). In Drosophila, Naa50 (also known as San) is required for establishing sister chromatid cohesion (Williams et al., 2003), something that has also been shown for NAA50 in human cell cultures (Hou et al., 2007). Further work in the fly suggests this phenotype occurs through defective interactions between cohesin subunits due to a lack of Nt-acetylation of the potential Naa50 substrate sister chromatid cohesion protein 1 (Scc1, also known as Vtd; see poster) (Ribeiro et al., 2016). Interestingly, co-depletion of NatA rescues the sister chromatid cohesion defects and the resulting mitotic arrest caused by NAA50 depletion in HeLa cells (Rong et al., 2016).

Regarding disease involvement, NAA50 has been implicated in cancer (see Box 1). Moreover, NAA50 has been identified as part of a co-expression network in the autoimmune inflammatory condition thyroid eye disease (Hu et al., 2022), and NAA50 expression is upregulated along with five other genes in the hippocampi of alcohol-treated female rats (Choi et al., 2020).

NatF comprises only the catalytic NAA60, which targets the same type of N-terminal sequences as NatC and NatE (Van Damme et al., 2011b). It is noteworthy, however, that NatF is the only one of these NATs that specifically targets membrane-bound proteins, and that NAA60 localizes to intracellular membranes (Aksnes et al., 2015). NatF may thus be referred to as a membrane NAT. Human NAA60 binds peripherally to the Golgi (Aksnes et al., 2017), where it is responsible for Nt-acetylation of membrane proteins with an N-terminal topology towards the cytosol (Aksnes et al., 2015). At this location, NAA60-mediated Nt-acetylation is likely to occur post- rather than co-translationally. The type of N termini preferred by NatF overlaps with those targeted by NatC and NatE – that is, sequences starting with a methionine, with the exception of those targeted by NatB. N-terminal proteomics of NAA60-depleted cells has identified 23 NatF substrate proteins (Aksnes et al., 2015). Crystal structure analyses have shown that the molecular interactions within the catalytic core of NAA60 are most similar to those in NAA50, and that NAA60 might form a homodimer (Stove et al., 2016; Liszczak et al., 2011; Chen et al., 2016).

Depletion of NAA60 causes delayed chromosome segregation during anaphase in Drosophila cells (Van Damme et al., 2011b) and fragmentation of the Golgi ribbon in HeLa cells (Aksnes et al., 2015) (see poster). NAA60 has recently been described as necessary for influenza A virus infection, since depletion and overexpression of NAA60 reduces and enhances, respectively, the virus growth via effects on IFNα signalling (Ahmed and Husain, 2021). The A. thaliana orthologue of NAA60 is vital for the high salt stress response and, contrary to its human counterpart, localizes to the plasma membrane together with one of its identified substrates (Linster et al., 2020).

Copy number variations in NAA60 have been identified in non-small cell lung cancer (Heo et al., 2021), but it appears to be less of a cancer-relevant gene than many of those encoding other NAT subunits (Koufaris and Kirmizis, 2020). NAA60 might be under epigenetic control, since it has been detected as differentially methylated or imprinted in various conditions, sometimes connected to disease (Kagami et al., 2017; Gu et al., 2020; Menzies et al., 2013; Nakabayashi et al., 2011).

NatH is another highly specific NAT, exclusively targeting actin proteins for Nt-acetylation (Drazic et al., 2018; Goris et al., 2018; Wiame et al., 2018) (see poster). The catalytic unit of NatH, NAA80, is involved in a distinct stepwise maturation process of the N termini of animal actins, following the actions of a co-translational NAT and a specific actin-maturation protease, ACTMAP (Haahr et al., 2022; Aksnes et al., 2019; Arnesen and Aksnes, 2023). NAA80 also associates with the actin-binding and -sequestering profilins, preferring PFN2 over the more abundant PFN1 (Rebowski et al., 2020; Ree et al., 2020). PFN2 may be considered as a key cofactor of NatH, as binding to PFN2 increases NAA80 enzymatic activity. Consistent with this, NAA80 does not associate with F-actin in cells, but rather favours G-actin bound to profilin (Rebowski et al., 2020).

The cellular roles defined for NatH are connected to functions of the actin cytoskeleton. NAA80-knockout cells have altered cytoskeletal organization, which is reflected in a decreased G-actin:F-actin ratio (Drazic et al., 2018) and increased F-actin staining intensity (Beigl et al., 2020). NAA80-knockout cells also exhibit more cytoskeletal protrusive structures and increased cell migration (Drazic et al., 2018), along with an increase in cell size and Golgi fragmentation (Beigl et al., 2020).

In agreement with the roles of NAA80 in regulation of the actin cytoskeleton, individuals harbouring a pathogenic NAA80 variant display decreased cellular actin Nt-acetylation and clinical features that overlap with those of pathogenic variants in actin genes (ACTG1, ACTB and ACTA1), such as high-frequency hearing loss, muscle weakness and developmental delay (Muffels et al., 2021).

Lines of evidence implicate Nt-acetylation in the regulation of protein folding, complex formation, subcellular targeting and lifetime (Aksnes et al., 2019). One of the first-proposed functions of the N-terminal acetyl group was protection from degradation (Hershko et al., 1984). Later, Nt-acetylation was shown to be part of conditional N-degrons that mediate protein degradation (Hwang et al., 2010; Shemorry et al., 2013). Recent data suggest that in human and plant cells, Nt-acetylation can confer stability to a number of modified proteins, and that a key role of Nt-acetylation might be protection from specific E3 ubiquitin ligases that recognize unmodified protein N termini (Mueller et al., 2021; Linster et al., 2022; Varland et al., 2022 preprint). More detailed studies in cellular models are anticipated to further define the potential of Nt-acetylation or lack thereof in directing the degradation of specific proteins.

Since the first reported cases just over a decade ago (Rope et al., 2011), several disease variants of NAT genes have been uncovered and described. Knowledge in this area is likely to grow to encompass additional pathological variants in NAT genes, thereby boosting NAT-related genetic diagnosis and helping to better elucidate the vital in vivo functions of this modification. A goal for future research will be to decipher the molecular pathways that are affected by lack of Nt-acetylation and any specific pathological outcomes. Moreover, it will be important to define more closely the specific involvement of NATs in various cancers (see Box 1) and the potential of targeting NATs in cancer treatment (see Box 2).

Nt-acetylation has in general been viewed as a constitutive and irreversible modification. This static view has, however, recently been challenged by the discovery of post-translational NATs, which appear more likely to be involved in Nt-acetylation regulated by an on–off switch, as compared to the co-translational Nt-acetylation during de novo synthesis of a protein. Furthermore, examples of responsive regulation have also been described among the co-translational NATs, including regulation of NatA in plant drought resistance (Linster et al., 2015), NatC in C. elegans stress resistance (Warnhoff et al., 2014) and NatD in yeast longevity (Molina-Serrano et al., 2016). Going forward, uncovering the pathways that act upstream of NATs to regulate their expression or activity, or that counteract NAT activity by catalysing deacetylation, will be an important research goal.

Finally, the question remains whether all NATs have been discovered. Although most of the Nt-acetylome appears to be accounted for by the known NATs, proteomics and phylogenetic data indicate that more NATs might exist. For example, no NAT has yet been shown to act in the mitochondrial matrix or inside any organellar lumen in human cells. Future discoveries of additional eukaryotic NAT enzymes are therefore not unlikely.

The authors are grateful to colleagues and collaborators for insightful discussions on various topics that are covered in this Cell Science at a Glance article.